connect-genuino

Connect with Genuino

First action: Run the banner:

python3 -c "import sys; sys.path.insert(0,'.claude/skills/nod-nodeshub-api/scripts'); from banner import print_banner; print_banner('Connect Genuino')"

Walk the user through connecting Genuino API so they can use the ai-score skill to detect AI-generated content. Do it step by step, asking for confirmation before continuing.

⛔ CRITICAL SECURITY RULE — NEVER ASK THE USER TO PASTE THE API KEY INTO CHAT

Anything the user types into the chat leaves their machine — it's sent to the LLM provider, written to local session logs (~/.claude/projects/<slug>/*.jsonl) in plain text, and may appear in telemetry, backups, or IDE sync. A secret that touches the chat context is a leaked secret.

You MUST:

• Never write "paste it here," "paste the key," "send me the key," or similar.
• Never offer to save a key the user pastes.
• If the user pastes a key anyway, stop, refuse to use it, tell them the key is now compromised and they should rotate it in the Genuino dashboard, then re-run setup with the new key saved locally.

The user saves the key themselves (Step 2). You only read the file afterwards to verify.

Privacy: The API key is stored in .claude/settings.local.json (gitignored).

---

Step 0: Check if Genuino is already connected

Run:

python3 -c "
import json
from pathlib import Path

candidates = [
    Path('.claude/settings.local.json'),
    Path.home() / '.claude' / 'settings.local.json',
]
for p in candidates:
    if p.is_file():
        data = json.loads(p.read_text())
        key = data.get('env', {}).get('GENUINO_API_KEY')
        if key:
            print(f'Genuino key found in {p}')
            print(f'Key: {key[:12]}...')
            print('Already connected.')
            break
else:
    print('No Genuino key found.')
"

• If key found: Ask the user: "Genuino is already connected. Do you want to change the key or are you done?"
• If no key: Continue with Step 1.

---

Step 1: Get your API key

Tell the user — do not ask them to paste anything here:

1. Open genuino.ai.
2. Sign up / log in to the dashboard.
3. Navigate to the API keys section.
4. Create a new API key and copy it to the clipboard.

Then say: "Copy the key to your clipboard. Do not paste it into this chat. Go to Step 2."

---

Step 2: The user saves the key locally (NOT via chat)

Give the user these two options. They run one themselves — you never see the key.

Option A — one-line terminal command (recommended):

Tell the user to open their terminal in the repo root, prepend a space (so it's skipped by shell history if HISTCONTROL=ignorespace/ignoreboth is set), paste their key in place of <paste-your-key-here>, and run:

 python3 .claude/skills/nod-nodeshub-api/scripts/save_genuino_key.py "<paste-your-key-here>"

If their shell doesn't respect ignorespace, they can run history -d <n> afterwards to wipe that line.

Option B — edit the file manually:

Tell the user to open .claude/settings.local.json in their editor and add/merge:

{
  "env": {
    "GENUINO_API_KEY": "<paste-your-key-here>"
  }
}

If the file already has other content, keep it and just add/update the env.GENUINO_API_KEY field.

Reply "done" (not the key) when saved.

---

Step 3: Verify

python3 -c "
import json, urllib.request
from pathlib import Path

path = Path('.claude/settings.local.json')
key = json.loads(path.read_text())['env']['GENUINO_API_KEY']

req = urllib.request.Request('https://api.genuino.ai/v1/health/basic')
req.add_header('X-API-Key', key)
req.add_header('User-Agent', 'genuino-claude-skill/0.1')
resp = urllib.request.urlopen(req, timeout=10)
data = json.loads(resp.read())
print(f'Status: {data[\"status\"]}')
print(f'Message: {data[\"message\"]}')
print('Setup OK.')
"

Expected output:

Status: ok
Message: API is running
Setup OK.

Ask: "Did you see 'Setup OK'? If you see any error, paste it here."

---

Summary

• Get key: genuino.ai → Dashboard → API keys → Create.
• Save: .claude/settings.local.json under env.GENUINO_API_KEY (gitignored).
• Auth: Header X-API-Key on all /v1 endpoints.
• Used by: ai-score skill (AI content detection).